Program to generate VE equations

[Bluecat]

I had made it until last month before I had to tune a C6 with a big cam. I had it in my mind for months how I was going to deal with the "non - ve table" stuff, so a few days before hand I sat down and wrote a program to generate equations from a ve talbe and the reverse. Well any way, it didn't work. at least not on the car. It generated numbers that made a pretty graph in a speadsheet, but the pcm had the airflow numbers boucing all over the place, even in the middle of a zone. I think I have it fixed now, unless there is some min or max values for the variables that I don't know about. But I don't have a car local with that style pcm to play with.

So here it is. Give me some feed back here on if the pcm will accept the numbers it gens or if it does a good job of matching user generated ve tables.

Its pretty simple. Data is kept in tab delimited text files that you can either edit manually or open in excel, just remeber to make excel save it as a tab delimeted text file. You can see the format with the supplied test files.



You feed it the equation variables and it spits out a ve table.

You feed it a ve table and it spits out the equation variables.

It has the ability to do either ls1 or ls2 format ve tables. Also VE numbers (you supply the clyinder volume in the equation.txt) or the pcm raw airflow format. Just rember that regaudless of importing or exporting ve tables, that you check the appropiate check box, because the check boxes apply to both in and out on the ve tables.

Pretty strait forward. Take the variables and make a ve table. Go log the car do what you would normally do with a ve table. Take your new ve table and have it generate the new variables.

* Link to V1 has been removed. V2 can be found here... *

http://www.hptuners.com/forum/showth...421#post140421

[Bill@HPTuners]

Very cool, someone should give this a shot, either way I'll put it in the sticky so it doesnt get lost for now, Good Work!

[redhardsupra]

this is AWESOME! are you going to share the source code?

[BBA]

This is great, now I can finally see the changes I've made in VE style.

Thanks!

[Bill@HPTuners]

I'd love to hear some feedback as to how its working out for you guys so I know to keep this stickied

[SSUte01]

That is absolutely brilliant, havent had a chance to do some real world testing but messing around with test data it is genius. Hope it works as well in practice now

Well Done

[deanm11]

Thanks so much for doing this. I am trying to compare to the spreadsheet that was created by carneb using xlfit addin. In the spreadsheet, each zone is calculated completely independently. What are you doing at the boundaries, if anything ? What does the 'automatic zone boundaries' toggle do ?

[Bluecat]

If you don't check automatic zone boundries it just gens the number with the rpm and map boundries that are in the current Equation.txt. When checked it looks for abrupt changes that would be hard for the flow of the variables to match and sets up the boundries to prevent poor transitions between zones.

As far as at the zones edges, it tries to work a balance between making the best match for the whole zone and still keeping the edges so that there isn't any harsh ledges at the zone transitions.

Max RPM should be set to your rev limiter. It stops the calc and compare at that rpm. To make a better match at 8000 rpm it might have to decrease the match at 6000 to compromise. If your not worried about what happens above a certain rpm (can't go higher than the rev limiter) then you can make a better aproximation below that point.

[deanm11]

Makes perfect sense what you did- well done!

If you don't check automatic zone boundries it just gens the number with the rpm and map boundries that are in the current Equation.txt. When checked it looks for abrupt changes that would be hard for the flow of the variables to match and sets up the boundries to prevent poor transitions between zones.

As far as at the zones edges, it tries to work a balance between making the best match for the whole zone and still keeping the edges so that there isn't any harsh ledges at the zone transitions.

Max RPM should be set to your rev limiter. It stops the calc and compare at that rpm. To make a better match at 8000 rpm it might have to decrease the match at 6000 to compromise. If your not worried about what happens above a certain rpm (can't go higher than the rev limiter) then you can make a better aproximation below that point.

[9secZO6]

Someone is doing this right now but using efi live ....e-38 controller tuning for drivability (off-idle, 1600rpm cruise etc.). Same thing, guy cam up w/ program that allows u to "back in" to a VE table. When I saw it in action I thought........now we are getting somewhere

[Bluecat]

Anyone get a chance to try and plug the numbers in a car yet?

[carneb]

This looks really good Bluecat. I've only had a quick play with it but it seems to get the job done. I don't have my wideband installed in the car at the moment so can't test it in the car.

The only problem I had was when I first tried to run it I got an error message about MSVBVM50.DLL file missing. I had to download and install the MS Visual Basic runtime files.

[Bluecat]

You wouldn't even have to use a wideband. Just log the dynamic g/sec or g/cly, or better yet the new "GM Volumetric Efficiency (raw VE value)" pid. If a historgram of the new raw ve pid matchs the ve table you made the equation variables from, we're good to go.

Yeah the last time I worked as a IT / Programming professional was in 2000. My copy of VB is a little old...

[PurplePiss]

First off. Thanks. Second, I have no idea how to use this. Can someone create a writeup?

Marcin, is this doing what your complicated excel sheet was going to do?

[redhardsupra]

yea, kinda, except mine also does bias tuning at the same time.

[Bluecat]

As stated above, on the vehicals that don't have a normal VE table but instead have the equation generated VE, it enables you to view and modify them as a normal VE table. It dosen't tune the VE tables for you or any other table for that matter, it just inputs and outputs VE tables.

If you need to use it and understand why, it should be self explanitory. If you'll open up the "Equation.txt" and "Table.txt" in excel you can see the format for the files and it should make sence on how to get the information in and out of the program.

I wouldn't care to do a write up on how to use it, but lets see if it works first.

[deanm11]

It is not working for me. I had tune my parameter using carneb's spreadsheet and xlfit. (did realize today by testing this that my first 'row' of map boundaries aren't all 30 ('08 LS3 vette) as the spreadsheet is configured, so changed my tune).

Attached are the starting table (this was my last table after applying % tweaks-my final effective table is just slightly different in spots due to the approximation of the parameters by xlfit), equations as calculated by the bluecat software and the resultant table from the bluecat parameters. Think there are some bugs to work out- take a look at the choppiness, ex., the 43 MAP column. I appreciate the effort and hope this helps.

Regards, Dean

[Bluecat]

It won't work with the table.txt like you have. It has to have the ve table in the standard ls1 or ls2 format to import. Those are the only two formats that the scanner will build a histogram in, so those are the only two I planned for.

Don't start with the VE table, start with the boundries, constant, map2, map,...ect that are currently in your pcm. Put them in "Equation.txt" along with the cylinder volume of your engine if you want to work with ve% instead of raw numbers (100% vs. 2671)

Let it build the VE table ("Gen Table" button) your going to modify in the ls2 format (don't check the ls1 box) since that's how the vcm scanner will show the histograms in your log. This becomes "Table_out.txt"

Modify the "Table_out.txt" how ever you would if the car had a VE table. Try and use good tuning practices here and build a VE table that is both smooth and realistic looking. If you feed my program something that looks like the rocky mountains you won't get good results. There is only so much it can do within the limits of the formulas. After your done, save it as "Table.txt"

Go back in and press the "Gen Equation" button and let it make the new VE Coefficients. You can do a visual compare of the green graph to the black one to see how close it approximated the VE table you gave it. If it looks like a good match, copy the data from "Equation_out.txt" back into your tune and see if it did what you wanted.

If you tried to subtract 15% fuel in the area of 800rpm/60Kpa, then hopefully on the next logging of the car it will reflect the changes. There are somethings I am assuming, one of which is the fact that the tables have more than one data set (DOD, Manifold switch,... ect) and that everyone knows to paste the new data into every category unless you were just specifically trying to tune for just one. Same goes for copying out, but since in the tunes I have had they were all set the same, I figured it would be common sence to make them all the same again when done.

[deanm11]

Do you mean the format of the presentation of my table.txt is wrong? You agree that the units in the file are ok? They are GMVE ("ls2 format").

It won't work with the table.txt like you have. It has to have the ve table in the standard ls1 or ls2 format to import. Those are the only two formats that the scanner will build a histogram in, so those are the only two I planned for.

Don't start with the VE table, start with the boundries, constant, map2, map,...ect that are currently in your pcm. Put them in "Equation.txt" along with the cylinder volume of your engine if you want to work with ve% instead of raw numbers (100% vs. 2671)

Let it build the VE table ("Gen Table" button) your going to modify in the ls2 format (don't check the ls1 box) since that's how the vcm scanner will show the histograms in your log. This becomes "Table_out.txt"

Modify the "Table_out.txt" how ever you would if the car had a VE table. Try and use good tuning practices here and build a VE table that is both smooth and realistic looking. If you feed my program something that looks like the rocky mountains you won't get good results. There is only so much it can do within the limits of the formulas. After your done, save it as "Table.txt"

Go back in and press the "Gen Equation" button and let it make the new VE Coefficients. You can do a visual compare of the green graph to the black one to see how close it approximated the VE table you gave it. If it looks like a good match, copy the data from "Equation_out.txt" back into your tune and see if it did what you wanted.

If you tried to subtract 15% fuel in the area of 800rpm/60Kpa, then hopefully on the next logging of the car it will reflect the changes. There are somethings I am assuming, one of which is the fact that the tables have more than one data set (DOD, Manifold switch,... ect) and that everyone knows to paste the new data into every category unless you were just specifically trying to tune for just one. Same goes for copying out, but since in the tunes I have had they were all set the same, I figured it would be common sence to make them all the same again when done.

[Bluecat]

The units are fine, but the standard LS2 VE table is a 17x17 grid with the RPM from 400 to 6800 in 400 rpm increments and the MAP from 15 to 105 in 5.625 KPA increments. If you have any questions just checkout some of the table.txt's. You can see LS1's and LS2's not only use different numbers, but the axis are swapped.

Since the Non-Table format dosen't actually have any standard boundries (the formulas would yield a number for 10000rpm and 200kpa if thats what you plug in), thier is no true format. But since HPtuners kept the histograms for these type vehicals in the last LS2 format used with a VE table, thats the way I support it. Unless you use the same format as the histograms in the scanner, the histograms are kinda useless.